Congratulations to professor of civil and environmental engineering David Sedlak for winning the Clarke Prize for excellence in water research thanks to his work in water reuse and reducing contaminants in water. Professor Sedlak was also in the news recently for discussing his belief that we are on the verge of our fourth revolution in providing water to the masses.
Professor Irina Conboy of bioengineering has found that oxytoxin—the hormone associated with sex, maternal instincts, trust, and romance—also plays a role in how younger people are healthier and less susceptible to cancer than our older and wrinkle-ier counterparts. And let's not forget last week's developments on how cellular environment also affects cellular health in older patients.
You may have seen videos of a new fleet of robots that are capable of locomotion or even running by modelling the mechanics of animals. This is cool and all, but what happens when a robot—which is incapable of assessing its environment—goes charging into icy terrain at 15 mph? It'd make a sweet YouTube video, but it would be a huge loss of all the resources that went into that robot's development.
Enter the collaboration between ETH Zurich and UC Berkeley, which proposes partnering up a super-expensive and powerful bot with a team of scouts, programmed with the human-like mission to work together to test the environment and help guide the more important machine. This coming together of organic and machine is like that adorable animal story where, against all odds, the puppy befriends the penguin and they just hang out together playing catch and sharing dinner.
[In the video below,] the big [robot] is ETH Zurich's StarlETH quadruped, while the little one is VelociRoACH, from UC Berkeley. VelociRoACH's job is to scurry around ahead of StarlETH, exploring the terrain. StarlETH watches the optical tag on VelociRoACH's back to see how the little robot is doing, and VelociRoACH also sends back IMU data.
Together, the robots are able to classify terrain as either slippery or safe to walk on with an accuracy of over 90 percent, and since StarlETH is able to localize VelociRoACH as it scampers around, StarlETH knows exactly where it is (or isn't) safe to step.
The risk here is that your poor little scout robot might end up stumbling onto a piece of terrain that's so dangerous that it gets stuck. This is better than your primary robot getting stuck (and the scout robot is still accomplishing what it was intended to accomplish by warning the primary robot of the danger), but you need to be prepared to consider these scout robots as expendable, which is why VelociRoACH is the robot of choice, as it's made mostly of cardboard, and you can just toss a bunch more into the mix to replace any robots that you may lose.
Wait a minute... So you're telling me the big robot callously sends forth a squad of expendable little robots to their doom and the little robots follow their orders mindlessly without the ability to question the scenario? This is actually sounding pretty robotic on second thought.
Birds and the Bees and the Bears
What happens when you ask Berkeley scientists about sex? They turn to their computers.
Some Berkeley researchers led by Professor Christos Papadimitriou pondered what they described as an evolutionary paradox: if evolution has created the perfect organism, then sexual reproduction will result in a loss of perfection in its offpring by diluting that ideal individual (e.g., me) with the genes of an inferior individual. The main thrust of this work was to answer the question: why do we have sexual reproduction?
They answered this question by partnering up just like you would when it's time to do the ole backside attack. Computer science collided with biology and game theory met evolutionary biology to break down the statistics and probability of mating.
First of all, they defined their problem. They decided not to look at life-or-death traits (like having gills when you live underwater), but rather "weak selection" where one trait is slightly preferential to another, like the gorgeous smile of a George Clooney vs. the stunning smile of yours truly; neither is decisively better.
"We noticed that with variation, genes have a preference for a 50-50 distribution rather than a 90-10 distribution," said Papadimitriou, a giant in the field of computational complexity. "If we use a gambling analogy, genes don't want to go all-in. They want to hedge their bets. Even if there is an extremely successful genetic trait, evolution doesn't want to let the genes for the other traits go extinct in case they're needed later."
While the genetic success of any random individual seems fleeting in this framework, the entire mix of genes gets better over time.
"Because genes are mixing so quickly, you can't think of evolution as acting on individuals," said Vazirani. "You must think of a soup consisting of genes of all individuals in a particular species. Evolution makes that soup better and better over time, regardless of what happens to any individual ingredient."
The key computational tool used to come to this conclusion is known as the multiplicative weight update algorithm, which is used to maximize profits from stocks. Rather than just buying one company's stock and trying to ride it all the way to your dreams of backstrokes in a gold-coin swimming pool, it diversifies your portfolio and basically reliably ensures profit overall.
So, let it be known—the world's leading experts in sex come from Berkeley. [NSFW] on you Bears?